I wish ASAN understood memory page boundaries and masked reads, because it sucks getting people telling you your code fails under ASAN when in fact it is correct code and you have to explain that it is that way on purpose :(
I am saying something significantly stronger: a tool that can't tell the difference between working code and non-working code just isn't useful unless you are willing to write slow code. So the correct way of saying it would be, "ASAN helps you find errors in slow code only."
-
-
That _should_ be considered unacceptable, in my opinion, for any project where speed is a concern. So if you were, for example, caring about the speed of your hash function in the first place, you should be considering ASAN not ready for prime time.
-
In the future, masked loads that don't fault off the page will be literally everywhere - it's literally the instruction everyone will have to call when doing loads on AVX-512. So what then?
- Show replies
New conversation -
-
-
I think the overwhelming majority of code is slow code (that's fine), and the overwhelming majority of fast code does not spend its time mostly doing partially out-of-bound reads.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.